A motion estimation method and device are provided for processing images to be inserted, between a preceding original image and a following original image, into a sequence of images. Each image is divided into pixel blocks associated with motion vectors. For a current block of an image being processed, motion vectors associated with blocks of the image being processed and/or associated with blocks of a processed image are selected. Candidate vectors are generated from selected motion vectors. An error is calculated for each candidate vector. A penalty is determined for a subset of candidate vectors on the basis of the values of the pixels of the pixel block in the preceding original image from which the candidate motion vector points to the current block and/or on the basis of the values of the pixels of the pixel block in the following original image to which the candidate motion vector points from the current block.
Legal claims defining the scope of protection, as filed with the USPTO.
1. A motion estimation method for processing images intended to be inserted, between a preceding original image and a following original image, into a sequence of images, each of the images being divided into a plurality of pixel blocks, and a motion vector being associated with each of the pixel blocks of a processed image, the method comprising, for a current block of an image being processed, the steps of: selecting spatial motion vectors associated with pixel blocks of the image being processed and/or temporal motion vectors associated with determined pixel blocks of an already processed image; generating candidate motion vectors from the selected motion vectors; calculating, by a processing device, an error for each of the candidate motion vectors; determining a penalty for at least one subset of the candidate motion vectors; electing a motion vector from among the candidate motion vectors based on the error that is calculated and the penalty that is determined; and storing information associating the elected motion vector with the current pixel block, wherein the determining step comprises, for each candidate motion vector of the subset: calculating a first value corresponding to a value of the pixels of the pixel block in the preceding original image from which that candidate motion vector points to the current block; calculating a second value corresponding to a value of the pixels of the pixel block in the preceding original image from which that candidate motion vector points to the pixel block with which the motion vector is associated, from which the candidate motion vector is generated, is associated; and comparing the first and second values to determine the penalty with a magnitude corresponding to a difference between the compared first and second values.
2. The method according to claim 1 , wherein determining further comprises calculating the penalty based on the comparing of the first and second values in such a way that the penalty is relatively large when the first and second values are substantially different and is relatively small when the first and second values are substantially equivalent.
3. The method according to claim 1 , wherein the determining step comprises, for a candidate motion vector generated from a temporal motion vector, the sub-steps of: calculating a third value corresponding to another value between pixels of the pixel block in the preceding original image from which the candidate vector points to the current pixel block; and determining the penalty based on the third value in such a way that the penalty is relatively large when the third value is relatively low and is relatively small when the third value is relatively high.
4. The method according to claim 3 , wherein the one value is one of a luminance value and a contrast value, and the other value is the other of the luminance value and the contrast value.
5. The method according to claim 1 , wherein the one value is one of a luminance value and a contrast value.
6. The method according to claim 1 , wherein the determining step comprises, for a candidate motion vector generated from a spatial motion vector, the sub-steps of: calculating a third value corresponding to one value of the pixels of the pixel block in the following original image to which the candidate motion vector points from the current block; calculating a fourth value corresponding to the one value of the pixels of the pixel block in the following original image to which the candidate motion vector points from the pixel block with which the motion vector, from which the candidate motion vector is generated, is associated; and determining the penalty based on a comparison between the third and fourth values in such a way that the penalty is relatively large when the third and fourth values are substantially different and is relatively small when the third and fourth values are substantially equivalent.
7. The method according to claim 1 , wherein the subset of candidate motion vectors includes the candidate motion vector having the lowest error and a determined number of candidate motion vectors having errors such that the difference between each of the errors and the lowest error is less than a determined threshold value.
8. The method according to claim 1 , wherein, for each candidate motion vector of the subset, the determining step comprises at least one of: calculating a first value corresponding to one value of the pixels of one of the pixel blocks in the preceding original image; and calculating a first value corresponding to the one value of the pixels of one of the pixel blocks in the following original image.
9. The method according to claim 1 , wherein, for each candidate motion vector of the subset, the determining step comprises determining the penalty based on one value of the pixels of that pixel block in the preceding original image from which that candidate motion vector points to the current block and/or based on the one value of the pixels of that pixel block in the following original image to which that candidate motion vector points from the current block.
10. A motion estimation device for processing images intended to be inserted, between a preceding original image and a following original image, into a sequence of images, each of the images being divided into a plurality of pixel blocks, and a motion vector being associated with each of the pixel blocks of a processed image, the device comprising, with respect to a current block of an image being processed: a selection unit for selecting spatial motion vectors associated with pixel blocks of the image being processed and/or temporal motion vectors associated with pixel blocks of an already processed image; a generation unit for generating candidate motion vectors from the selected motion vectors; a calculation unit for calculating an error for each of the candidate motion vectors; a determination unit for determining a penalty for at least one subset of the candidate motion vectors; and an election unit for electing a motion vector from among the candidate motion vectors based on the error that is calculated and the penalty that is determined; and a memory for storing information associating the elected motion vector with the current pixel block; wherein the determination unit comprises: a first calculation unit for calculating a first value corresponding to one value of the pixels of the pixel block in the preceding original image from which the candidate motion vector points to the current block; and a second calculation unit for calculating a second value corresponding to the one value of the pixels of the pixel block in the preceding original image from which the candidate motion vector points to the pixel block with which the motion vector, from which the candidate motion vector is generated, is associated; said determination unit determines the penalty based on a comparison of the first and second values to have a magnitude corresponding to a difference between the compared first and second values.
11. The device according to claim 10 , wherein the determination unit determines the penalty based on comparing the first and second values so that the penalty is relatively large when the first and second values are substantially different and is relatively small when the first and second values are substantially equivalent.
12. The device according to claim 11 , further comprising, for a candidate motion vector generated from a spatial motion vector: a third calculation unit for calculating a third value corresponding to the one value of the pixels of the pixel block in the following original image to which the candidate motion vector points from the current block; and a fourth calculation unit for calculating a fourth value corresponding to the one value of the pixels of the pixel block in the following original image to which the candidate motion vector points from the pixel block with which the motion vector, from which the candidate motion vector is generated, is associated, wherein the determination unit determines the penalty based on a comparison between the third and fourth values so that the penalty is relatively large when the third and fourth values are substantially different and is relatively small when the third and fourth values are substantially equivalent.
13. The device according to claim 12 , furthermore comprising, for a candidate motion vector generated from a temporal motion vector: a fifth calculation unit for calculating a fifth value corresponding to another value between pixels of the pixel block of the preceding original image from which the candidate vector points to the current pixel block, wherein the determination unit determines the penalty based on the fifth value, so that the penalty is relatively large when the fifth value is relatively low and is relatively small when the fifth value is relatively high.
14. The device according to claim 10 , wherein the subset of candidate motion vectors includes the candidate motion vector having the lowest error and a determined number of candidate motion vectors having errors, such that the difference between each of the errors and the lowest error is less than a determined threshold value.
15. A non-transitory computer readable medium encoded with a program for processing images intended to be inserted, between a preceding original image and a following original image, into a sequence of images, each of the images being divided into a plurality of pixel blocks, and a motion vector being associated with each of the pixel blocks of a processed image, the program including instructions for performing, for a current block of an image being processed, the steps of: selecting spatial motion vectors associated with pixel blocks of the image being processed and/or temporal motion vectors associated with determined pixel blocks of an already processed image; generating candidate motion vectors from the selected motion vectors; calculating an error for each of the candidate motion vectors; determining a penalty for at least one subset of the candidate motion vectors; electing a motion vector from among the candidate motion vectors based on the error that is calculated and the penalty that is determined; and storing information associating the elected motion vector with the current pixel block, wherein determining comprises: calculating a first value corresponding to one value of the pixels of the pixel block in the preceding original image from which the candidate motion vector points to the current block; calculating a second value corresponding to the one value of the pixels of the pixel block in the preceding original image from which the candidate motion vector points to the pixel block with which the motion vector is associated, from which the candidate motion vector is generated; and comparing the first and second values to determine the penalty with a magnitude corresponding to a difference between the compared first and second values.
16. The non-transitory computer readable medium according to claim 15 , wherein determining comprises calculating the penalty based on the comparing of the first and second values in such a way that the penalty is relatively large when the first and second values are substantially different and is relatively small when the first and second values are substantially equivalent.
17. The non-transitory computer readable medium according to claim 15 , wherein the subset of candidate motion vectors includes the candidate motion vector having the lowest error and a determined number of candidate motion vectors having errors such that the difference between each of the errors and the lowest error is less than a determined threshold value.
18. A motion estimation method for processing images intended to be inserted, between a preceding original image and a following original image, into a sequence of images, each of the images being divided into a plurality of pixel blocks, and a motion vector being associated with each of the pixel blocks of a processed image, the method comprising, for a current block of an image being processed, the steps of: selecting spatial motion vectors associated with pixel blocks of the image being processed and/or temporal motion vectors associated with determined pixel blocks of an already processed image; generating candidate motion vectors from the selected motion vectors; calculating, by a processing device, an error for each of the candidate motion vectors; determining a penalty for at least one subset of the candidate motion vectors; electing a motion vector from among the candidate motion vectors based on the error that is calculated and the penalty that is determined; and storing information associating the elected motion vector with the current pixel block, wherein the determining step comprises, for each candidate motion vector of the subset: calculating a first value corresponding to the one value of the pixels of the pixel block in the following original image to which the candidate motion vector points from the current block; calculating a second value corresponding to the one value of the pixels of the pixel block in the following original image to which the candidate motion vector points from the pixel block with which the motion vector, from which the candidate motion vector is generated, is associated; and comparing the first and second values to determine the penalty with a magnitude corresponding to a difference between the compared first and second values.
19. The method according to claim 18 , wherein determining further comprises calculating the penalty based on the comparing of the third and fourth values in such a way that the penalty is relatively large when the first and second values are substantially different and is relatively small when the first and second values are substantially equivalent.
20. The method according to claim 19 , wherein the determining step further comprises, for a candidate motion vector generated from a temporal motion vector, the sub-steps of: calculating a third value corresponding to another value between pixels of the pixel block in the preceding original image from which the candidate vector points to the current pixel block; and determining the penalty based on the third value in such a way that the penalty is relatively large when the third value is relatively low and is relatively small when the third value is relatively high.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
January 13, 2014
March 24, 2015
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.